Naked eye calculations:
Make Vultra=10V (instead of 5V)
Choose:
Ic=15mA (for lowest noise)
Vce=4.2V (for max Vce excursion)
-----------
Re=100ohm (default input impedance)
Rc=270ohm (for a decent gain)
Base divider:
R9=10k + 2.7k to ground (for ~15mA Ic)
Now, gain is ~Rc/Rs= 27 (for Rs=10ohm) and OLM is (for 0.5mV nominal @1KHz) almost 30dB (4.2V in Vce excursion divided by 135mV nominal from cartridge gain excursion @20KHz).
Input impedance will be about 60ohm due to the low input impedance of the BJT (Beta/gm), at 15mA.
Done.
Make Vultra=10V (instead of 5V)
Choose:
Ic=15mA (for lowest noise)
Vce=4.2V (for max Vce excursion)
-----------
Re=100ohm (default input impedance)
Rc=270ohm (for a decent gain)
Base divider:
R9=10k + 2.7k to ground (for ~15mA Ic)
Now, gain is ~Rc/Rs= 27 (for Rs=10ohm) and OLM is (for 0.5mV nominal @1KHz) almost 30dB (4.2V in Vce excursion divided by 135mV nominal from cartridge gain excursion @20KHz).
Input impedance will be about 60ohm due to the low input impedance of the BJT (Beta/gm), at 15mA.
Done.
I would leave input stage with a fixed gain, and the diff to single ended stage with unity gain and deal with the total gain in the RIAA stage(s), I assume you will go with active filters.
BTW, expect noise to be 3dB higher than a single transistor, due to uncorrelated noise sources in the differential stage.
What about people using their existing MM amp without selectable gain ?
A non variable gain head amp where gain will depend on the used Cart is less universal.
Changing Ic from 9.3mA to 15mA will only drop EIN from 0.31 to 0.28nV/rtHz.
Is that worth the trouble with the LM4562 that can no longer supply this current?
Raising the 5Volt supply would help to increase OLM to a higher level although without feedback in the first stage this amp is IMO already less sensitive to the effects of getting overloaded, and will with the current OLM prevent the Riaa stage from getting overloaded, agree ?
Hans
Let me sim these proposals later today and see what we can get. I can always parallel two LM4562 or add a small pass transistor for the 15 mA per side Ic.
Agree noise on this design won’t be as quiet - but it’s still pretty good IMV plus you get the CMR benefit
Agree noise on this design won’t be as quiet - but it’s still pretty good IMV plus you get the CMR benefit
Last edited:
I don’t disagree with anything you are saying, only that it started to sound like a project in search of requirements 😀.
We used to call this ‘creeping design’. It’s why in auto development they have a hard stop in the spec. On a specific date, no more changes to the spec - It’s frozen and the program moves onto implementation.
To further illustrate the point I mentioned with overloading the first stage having no feedback, the image below shows the output signal for the most critical 1R Cart with 50mV rms input at the left.
This is the 20dB OLM that I gave in the earlier graph.
At the right is shown what happens with 30dB overload or 158mV rms input.
As can be seen, no drama at all with smooth soft clipping.
Hans
.
This is the 20dB OLM that I gave in the earlier graph.
At the right is shown what happens with 30dB overload or 158mV rms input.
As can be seen, no drama at all with smooth soft clipping.
Hans
.
Attachments
I've done more work on this tonight. Front end supply raised to 10V and the collector currents in each half set to 15mA.
The output current of a number of cartridges is shown in the table below. The output range for these carts is c. 12uA up to 63uA covering Denon DL103 to Goldring Ethos. These figures include the amplifier Rin error.
The absolute worst case overload situation is with the Goldring Ethos. At 26 dB overload at 20 kHz this corresponds to 13.5mA out of the cart. Not likely in practice but it is what it is.
I am going re-arrange the collector load switching to cover 40 to 250 Ohms nominally providing 5mV out of the head amp for the range of cartridge currents above.
Distortion at +26 dB OL on the Goldring Ethos at 20 kHz (13.5mA input) is 0.15% It drops off very quickly below this at at 20 kHz levels reference the nominal 1kHz level is < 20 ppm.
1kHz spot noise is 375pv/rt Hz, 53nV RMS
I will update the circuit tomorrow.
The output current of a number of cartridges is shown in the table below. The output range for these carts is c. 12uA up to 63uA covering Denon DL103 to Goldring Ethos. These figures include the amplifier Rin error.
The absolute worst case overload situation is with the Goldring Ethos. At 26 dB overload at 20 kHz this corresponds to 13.5mA out of the cart. Not likely in practice but it is what it is.
I am going re-arrange the collector load switching to cover 40 to 250 Ohms nominally providing 5mV out of the head amp for the range of cartridge currents above.
Distortion at +26 dB OL on the Goldring Ethos at 20 kHz (13.5mA input) is 0.15% It drops off very quickly below this at at 20 kHz levels reference the nominal 1kHz level is < 20 ppm.
1kHz spot noise is 375pv/rt Hz, 53nV RMS
I will update the circuit tomorrow.
Attachments
Last edited:
For the DIY audio crowd, I believe you could skip a few lines from the table. Like the Ortofon MC Anna, $10k, I'm myself not that crazy, unless the latest version is able to clean the house, cook, take the garbage out, and perform in the bedroom.
Just another idea:
It would be possible to supply the whole circuit with one single 5volt supply when using an external USB power pack that delivers impeccable noise free supply.
This will save one or two LM4562's, but instead an output cap and an easy to use on board micro USB connector.
With the proper RRIO amps, performance will be unchanged and gain up to 30dB still possible.
Hans
.
It would be possible to supply the whole circuit with one single 5volt supply when using an external USB power pack that delivers impeccable noise free supply.
This will save one or two LM4562's, but instead an output cap and an easy to use on board micro USB connector.
With the proper RRIO amps, performance will be unchanged and gain up to 30dB still possible.
Hans
.
Attachments
Hans,
we increased the supply to the MC Head amp to 10V to ensure 26 dB OLM at max possible MC cart output (Goldring Ethos) and also to lower the thermal noise. Let me lay the board out for the complete thing, and then I will strip the MC head amp part out as a separate board that can then be configured for 5V USB supply Head amp only.
Final circuit diagrams coming now in the next posts. As always, comments welcome.
we increased the supply to the MC Head amp to 10V to ensure 26 dB OLM at max possible MC cart output (Goldring Ethos) and also to lower the thermal noise. Let me lay the board out for the complete thing, and then I will strip the MC head amp part out as a separate board that can then be configured for 5V USB supply Head amp only.
Final circuit diagrams coming now in the next posts. As always, comments welcome.
Here is the MC head amp circuit. Updates from previous version:
10V supply
Improved supply (higher O/P current
Load switched from 50-250 Ohms in 50 Ohm increments to maximize dynamic range (vs fixed 300 Ohm load previously that was tapped off - this limited the OLM on high current output carts)
10V supply
Improved supply (higher O/P current
Load switched from 50-250 Ohms in 50 Ohm increments to maximize dynamic range (vs fixed 300 Ohm load previously that was tapped off - this limited the OLM on high current output carts)
Attachments
This is the MM RIAA EQ stage. It incorporates MM cart resistive loading, a system gain stage (1x, 2x, 3x and 4x switchable gain) and a 20 Hz 60 db/decade HPF.
Power supply to the complete board is +-15V at c. 80 mA.
The unit is housed in the same type of box as the X-Altra (Modushop) and the rear panel is also a PCB pre routed/drilled and silk screened.
The +-15V supply is housed remotely and connected to the main amp board via a 3 core XLR cable 1 to 3 metres long.
Power supply to the complete board is +-15V at c. 80 mA.
The unit is housed in the same type of box as the X-Altra (Modushop) and the rear panel is also a PCB pre routed/drilled and silk screened.
The +-15V supply is housed remotely and connected to the main amp board via a 3 core XLR cable 1 to 3 metres long.
Attachments
Last edited:
I would make the rumble filter switchable. Also, I see you included the Neumann pole, I would make that configurable too (DIP switch), some audiophiles find this pole highly objectionable (no idea why).
P.S. My preference is to have the 75uS and 750uS poles in separate stages (mostly for accuracy reasons, I was able to bring it in +/-0.1dB). Then you could add or remove gain in the LPF stage (first after the pre-pre, to avoid any downstream OLM), so no extra op amp. Not that it really matters...
Last edited: